CN110570244A - hot-selling commodity construction method and system based on abnormal user identification - Google Patents
hot-selling commodity construction method and system based on abnormal user identification Download PDFInfo
- Publication number
- CN110570244A CN110570244A CN201910829742.8A CN201910829742A CN110570244A CN 110570244 A CN110570244 A CN 110570244A CN 201910829742 A CN201910829742 A CN 201910829742A CN 110570244 A CN110570244 A CN 110570244A
- Authority
- CN
- China
- Prior art keywords
- data
- user
- hot
- abnormal
- abnormal user
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- General Engineering & Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Economics (AREA)
- Evolutionary Computation (AREA)
- Game Theory and Decision Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a hot commodity construction method and system based on abnormal user identification, wherein the method comprises the following steps: extracting features from the user behavior sequence data, constructing a user feature vector to generate training data, and then training an isolated forest model by using the training data to identify abnormal users; marking the behavior data generated by the abnormal user in the test log data based on the real-time detection result of the abnormal user of the message sequence, thereby identifying the abnormal user; when the hot-sold commodity pool is updated by using the statistical result to generate the hot-sold commodities, the data marked as abnormal users are discarded, and therefore the hot-sold pool list is constructed. The invention eliminates abnormal users, thereby constructing a more accurate hot-selling commodity pool and truly and effectively reflecting real hot-selling commodities.
Description
Technical Field
the patent application belongs to the technical field of e-commerce systems, and particularly relates to a hot-market commodity construction method and system based on abnormal user identification.
background
the commodity hot sales pool in the current e-commerce recommendation system is constructed and depends on statistical data such as commodity browsing amount, volume of transaction, conversion rate and the like. The technical disadvantages thus exposed are: the strategy for generating the hot-sold commodities according to the statistical data is easily subjected to targeted attacks (such as bill swiping, malicious comment, malicious rejection and the like) of abnormal users, so that the authenticity and the effectiveness of the hot-sold commodities are influenced, and malicious competition among merchants is aggravated.
disclosure of Invention
the invention provides a hot-selling commodity construction method and system based on abnormal user identification, which can truly and effectively reflect real hot-selling commodities.
In order to solve the problems, the technical scheme adopted by the invention is as follows:
A hot-sold commodity construction method based on abnormal user identification comprises the following steps:
A, extracting features from user behavior sequence data, constructing a user feature vector to generate training data, and then training an isolated forest model by using the training data to identify abnormal users;
b, marking the behavior data generated by the abnormal user in the test log data based on the real-time detection result of the abnormal user of the message sequence, thereby identifying the abnormal user;
And c, when the hot-sold commodity pool is updated by using the statistical result to generate the hot-sold commodities, discarding the data marked as the abnormal user, thereby constructing a hot-sold pool list.
the technical scheme of the invention is further improved as follows: in the step a, an isolated forest model is trained in a distributed mode, and the specific training process is as follows:
randomly dividing the generated training data into K parts, and training an isolated forest model on a pc machine by using the same training parameters for each part to respectively obtain an isolated forest model;
after training is finished, packaging the K isolated forest models together, each isolated forest model can score a user, the score is the probability that the user is an abnormal user, and the average value, the median value or the public value of the scores of the K isolated forest models is taken as a final prediction score.
The technical scheme of the invention is further improved as follows: the isolated forest model comprises m isolated trees, each isolated tree is of a binary tree structure, and the implementation steps of the isolated forest model are as follows: (assuming that the total number of training data is A, the data size of each isolated forest model is A/K/m)
step a1. randomly selecting x sample points from a piece of training data as downsamples, and putting the downsamples into a root node of a tree;
step a2, randomly appointing a dimension attribute, and randomly generating a cutting point p in the current node data, wherein the cutting point p is generated between the maximum value and the minimum value of the appointed dimension attribute in the current node data;
Step a3. generates a hyperplane with this cut point p, then divides the current node data space into 2 subspaces: placing data smaller than p in the specified dimension on the left child of the current node, and placing data larger than or equal to p on the right child of the current node;
Step a4. recurses steps a2 and a3 in the child nodes, and new child nodes are continuously constructed until only one piece of data in the child nodes can not be cut any more or the child nodes reach a limited height.
the technical scheme of the invention is further improved as follows: the specific process of identifying the abnormal user comprises the following steps:
for a test log data y, traversing each isolated tree, and then calculating the height value of y which finally falls on the fourth layer of each isolated tree, namely y is in the height value of the tree, so as to obtain the height average value of y on each tree; (i.e., the average path length over t iTrees)
after the average height value (the average length of the length over t i tress) of each test log data is obtained, the comparison is performed through a set boundary threshold, and if the average height value (the average length of the length over t i tress) is lower than the boundary threshold, the test data is abnormal (that is, the abnormal has only a short average height in the trees).
A hot-sell commodity construction system based on abnormal user identification is used for achieving the construction method and comprises an abnormal user detection system, a data identification system and a data cleaning system which are sequentially in communication connection, wherein the abnormal user detection system comprises a back-end server used for extracting user behavior data, a consumer click stream processing platform connected with a user behavior log formed by the back-end server, a user portrait data repository connected with the consumer click stream processing platform, and a model offline training module connected with the user portrait data repository, and the model offline training module is used for generating an isolated forest model;
the data identification system comprises an abnormal user data identification module connected with the isolated forest model;
the data cleaning system comprises a data cleaning module connected with the abnormal user data identification module and the commodity picture database, and a real commodity picture database connected with the data cleaning module, and the real commodity picture database is utilized to construct a commodity hot sales pool list.
The technical scheme of the invention is further improved as follows: the consumer click stream processing platform is kafka.
The technical scheme of the invention is further improved as follows: the specific working process of the data identification system is as follows:
1) extracting a feature vector describing the user based on the user historical behavior sequence data;
2) Using an isolated forest algorithm of unsupervised learning to train an isolated forest model in a distributed mode;
3) And after the isolated forest model is obtained through training, scoring the user, and marking the user as a normal user or an abnormal user according to a scoring result.
Due to the adoption of the technical scheme, the invention has the beneficial effects that:
The invention constructs an abnormal user detection system in an E-commerce recommendation project based on an isolated forest model, and the idea of the isolated forest model is as follows: assuming we cut (split) the data space (data space) with one random hyperplane, two subspaces can be generated by cutting once (imagine cutting a cake with a knife in half). We then continue to cut each subspace with a random hyperplane, looping on until there is only one data point inside each subspace. Intuitively, we can see that clusters with high density can be cut many times before cutting stops, but that points with low density can easily stop in a subspace early. The isolated forest is a tree model based on an ensemble idea, and has good generalization performance and interpretability on massive high-dimensional data in a recommendation system.
and marking the data generated by the legacy users in the log data based on the real-time detection result of the abnormal user detection system, so that the commodity hot sales list generated based on the statistical result has higher reliability by marking the data generated by the abnormal users.
When the hot-sell commodity pool is updated by using the statistical result, the data marked as the data generated by the abnormal user is discarded, so that more accurate data is obtained, and ordered and targeted competition is facilitated.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
Detailed Description
The present invention will be described in further detail with reference to examples.
the invention discloses a hot commodity construction method based on abnormal user identification, which comprises the following steps:
A, extracting features from user behavior sequence data, constructing a user feature vector to generate training data, and then training an isolated forest model by using the training data to identify abnormal users;
B, marking the behavior data generated by the abnormal user in the test log data based on the real-time detection result of the abnormal user of the message sequence, thereby identifying the abnormal user;
And c, when the hot-sold commodity pool is updated by using the statistical result to generate the hot-sold commodities, discarding the data marked as the abnormal user, thereby constructing a hot-sold pool list.
in the step a, an isolated forest model is trained in a distributed mode, and the specific training process is as follows:
Randomly dividing the generated training data into K parts, and training an isolated forest model on a pc machine by using the same training parameters for each part to respectively obtain an isolated forest model;
After training is finished, packaging the K isolated forest models together, each isolated forest model can score a user, the score is the probability that the user is an abnormal user, and the average value, the median value or the public value of the scores of the K isolated forest models is taken as a final prediction score.
The isolated forest model comprises m isolated trees, each isolated tree is of a binary tree structure, and the implementation steps of the isolated forest model are as follows: (assuming that the total number of training data is A, the data size of each isolated forest model is A/K/m)
Step a1. randomly selecting x sample points from a piece of training data as downsamples, and putting the downsamples into a root node of a tree;
step a2, randomly appointing a dimension attribute, and randomly generating a cutting point p in the current node data, wherein the cutting point p is generated between the maximum value and the minimum value of the appointed dimension attribute in the current node data;
step a3. generates a hyperplane with this cut point p, then divides the current node data space into 2 subspaces: placing data smaller than p in the specified dimension on the left child of the current node, and placing data larger than or equal to p on the right child of the current node;
step a4. recurses steps a2 and a3 in the child nodes, and new child nodes are continuously constructed until only one piece of data in the child nodes can not be cut any more or the child nodes reach a limited height.
the specific process of identifying the abnormal user comprises the following steps:
for a test log data y, traversing each isolated tree, and then calculating the height value of y which finally falls on the fourth layer of each isolated tree, namely y is in the height value of the tree, so as to obtain the height average value of y on each tree; (i.e., the average path length over t iTrees)
After the average height value (the average length of the length over t i tress) of each test log data is obtained, the comparison is performed through a set boundary threshold, and if the average height value (the average length of the length over t i tress) is lower than the boundary threshold, the test data is abnormal (that is, the abnormal has only a short average height in the trees).
a hot-sell commodity construction system based on abnormal user identification is used for achieving the construction method and comprises an abnormal user detection system, a data identification system and a data cleaning system which are sequentially in communication connection, wherein the abnormal user detection system comprises a back-end server used for extracting user behavior data, a consumer click stream processing platform connected with a user behavior log formed by the back-end server, a user portrait data repository connected with the consumer click stream processing platform, and a model offline training module connected with the user portrait data repository, and the model offline training module is used for generating an isolated forest model;
the data identification system comprises an abnormal user data identification module connected with the isolated forest model;
the data cleaning system comprises a data cleaning module connected with the abnormal user data identification module and the commodity picture database, and a real commodity picture database connected with the data cleaning module, and the real commodity picture database is utilized to construct a commodity hot sales pool list.
the consumer click stream processing platform is kafka.
The specific working process of the data identification system is as follows:
1) extracting a feature vector describing the user based on the user historical behavior sequence data;
2) Using an isolated forest algorithm of unsupervised learning to train an isolated forest model in a distributed mode;
3) And after the isolated forest model is obtained through training, scoring the user, and marking the user as a normal user or an abnormal user according to a scoring result.
the data volume in the recommendation system is large, the memory of a single machine is usually insufficient to train a complete isolated forest model, and the model can be trained in a distributed mode because the isolated forest is an ensemble model.
after generating the user's features based on the user behavior sequence data, training data is generated and randomly divided into K shares, each of which trains an isolated forest model on a pc machine (using the same training parameters).
After training is finished, packaging the K isolated forest models together, each model can score a user (the score can be understood as the probability that the user is an abnormal user), and taking the average value of the scores of the K models as a final prediction score.
When the method is used, the commodity hot sales pool in the e-commerce recommendation system can be constructed without depending on statistical data such as commodity browsing amount, transaction amount, conversion rate and the like, and after users are marked through the isolated forest model, dynamic data generated by abnormal users are removed, the abnormal users are removed, so that a more accurate hot sales commodity pool is constructed, and real hot sales commodities can be truly and effectively reflected.
Claims (7)
1. a hot-sell commodity construction method based on abnormal user identification is characterized by comprising the following steps:
a, extracting features from user behavior sequence data, constructing a user feature vector to generate training data, and then training an isolated forest model by using the training data to identify abnormal users;
B, marking the behavior data generated by the abnormal user in the test log data based on the real-time detection result of the abnormal user of the message sequence, thereby identifying the abnormal user;
And c, when the hot-sold commodity pool is updated by using the statistical result to generate the hot-sold commodities, discarding the data marked as the abnormal user, thereby constructing a hot-sold pool list.
2. the hot-sold commodity construction method based on abnormal user identification as claimed in claim 1, wherein: in the step a, an isolated forest model is trained in a distributed mode, and the specific training process is as follows:
Randomly dividing the generated training data into K parts, and training an isolated forest model on a pc machine by using the same training parameters for each part to respectively obtain an isolated forest model;
After training is finished, packaging the K isolated forest models together, each isolated forest model can score a user, the score is the probability that the user is an abnormal user, and the average value, the median value or the public value of the scores of the K isolated forest models is taken as a final prediction score.
3. the hot-sold commodity construction method based on abnormal user identification as claimed in claim 2, wherein: the isolated forest model comprises m isolated trees, each isolated tree is of a binary tree structure, and the implementation steps of the isolated forest model are as follows:
step a1. randomly selecting x sample points from a piece of training data as downsamples, and putting the downsamples into a root node of a tree;
Step a2, randomly appointing a dimension attribute, and randomly generating a cutting point p in the current node data, wherein the cutting point p is generated between the maximum value and the minimum value of the appointed dimension attribute in the current node data;
Step a3. generates a hyperplane with this cut point p, then divides the current node data space into 2 subspaces: placing data smaller than p in the specified dimension on the left child of the current node, and placing data larger than or equal to p on the right child of the current node;
Step a4. recurses steps a2 and a3 in the child nodes, and new child nodes are continuously constructed until only one piece of data in the child nodes can not be cut any more or the child nodes reach a limited height.
4. The hot-sell commodity construction method based on abnormal user identification according to claim 3, wherein: the specific process of identifying the abnormal user comprises the following steps:
For a test log data y, traversing each isolated tree, and then calculating the height value of y which finally falls on the fourth layer of each isolated tree, namely y is in the height value of the tree, so as to obtain the height average value of y on each tree;
after the height average value of each test log data is obtained, the comparison is performed through a set boundary threshold value, and if the height average value is lower than the boundary threshold value, the test data is an exception (that is, the exception has a short average height in the trees).
5. A hot commodity construction system based on abnormal user identification is characterized in that: the construction method comprises an abnormal user detection system, a data identification system and a data cleaning system which are sequentially in communication connection, wherein the abnormal user detection system comprises a rear-end server for extracting user behavior data, a consumer click stream processing platform connected with a user behavior log formed by the rear-end server, a user portrait data storage library connected with the consumer click stream processing platform, and a model offline training module connected with the user portrait data storage library, and the model offline training module is used for generating an isolated forest model;
the data identification system comprises an abnormal user data identification module connected with the isolated forest model;
the data cleaning system comprises a data cleaning module connected with the abnormal user data identification module and the commodity picture database, and a real commodity picture database connected with the data cleaning module, and the real commodity picture database is utilized to construct a commodity hot sales pool list.
6. The hot-sell commodity building system based on abnormal user identification according to claim 5, wherein: the consumer click stream processing platform is kafka.
7. the hot-sell commodity building system based on abnormal user identification according to claim 5, wherein: the specific working process of the data identification system is as follows:
1) extracting a feature vector describing the user based on the user historical behavior sequence data;
2) using an isolated forest algorithm of unsupervised learning to train an isolated forest model in a distributed mode;
3) and after the isolated forest model is obtained through training, scoring the user, and marking the user as a normal user or an abnormal user according to a scoring result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910829742.8A CN110570244A (en) | 2019-09-04 | 2019-09-04 | hot-selling commodity construction method and system based on abnormal user identification |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910829742.8A CN110570244A (en) | 2019-09-04 | 2019-09-04 | hot-selling commodity construction method and system based on abnormal user identification |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110570244A true CN110570244A (en) | 2019-12-13 |
Family
ID=68777479
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910829742.8A Pending CN110570244A (en) | 2019-09-04 | 2019-09-04 | hot-selling commodity construction method and system based on abnormal user identification |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110570244A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112738088A (en) * | 2020-12-28 | 2021-04-30 | 上海观安信息技术股份有限公司 | Behavior sequence anomaly detection method and system based on unsupervised algorithm |
CN113887674A (en) * | 2021-12-06 | 2022-01-04 | 深圳索信达数据技术有限公司 | Abnormal behavior detection method and system based on big data |
WO2022047659A1 (en) * | 2020-09-02 | 2022-03-10 | 大连大学 | Multi-source heterogeneous log analysis method |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TW201310377A (en) * | 2011-08-16 | 2013-03-01 | Alibaba Group Holding Ltd | Recommending content information based on user behavior |
CN103853948A (en) * | 2012-11-28 | 2014-06-11 | 阿里巴巴集团控股有限公司 | User identity recognizing and information filtering and searching method and server |
CN108108743A (en) * | 2016-11-24 | 2018-06-01 | 百度在线网络技术(北京)有限公司 | Abnormal user recognition methods and the device for identifying abnormal user |
CN108320171A (en) * | 2017-01-17 | 2018-07-24 | 北京京东尚科信息技术有限公司 | Hot item prediction technique, system and device |
KR101898622B1 (en) * | 2017-07-24 | 2018-09-13 | 남기준 | Method for providing dealing with problematic customer service |
-
2019
- 2019-09-04 CN CN201910829742.8A patent/CN110570244A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TW201310377A (en) * | 2011-08-16 | 2013-03-01 | Alibaba Group Holding Ltd | Recommending content information based on user behavior |
CN103853948A (en) * | 2012-11-28 | 2014-06-11 | 阿里巴巴集团控股有限公司 | User identity recognizing and information filtering and searching method and server |
CN108108743A (en) * | 2016-11-24 | 2018-06-01 | 百度在线网络技术(北京)有限公司 | Abnormal user recognition methods and the device for identifying abnormal user |
CN108320171A (en) * | 2017-01-17 | 2018-07-24 | 北京京东尚科信息技术有限公司 | Hot item prediction technique, system and device |
KR101898622B1 (en) * | 2017-07-24 | 2018-09-13 | 남기준 | Method for providing dealing with problematic customer service |
Non-Patent Citations (2)
Title |
---|
FEI TONY LIU 等: "Isolation-Based Anomaly Detection", 《ACM TRANSACTIONS ON KNOWLEDGE DISCOVERY FROM DATA》 * |
YEZHU: "iForest(Isolation Forest) 孤立森林异常检测入门篇", 《简书》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2022047659A1 (en) * | 2020-09-02 | 2022-03-10 | 大连大学 | Multi-source heterogeneous log analysis method |
CN112738088A (en) * | 2020-12-28 | 2021-04-30 | 上海观安信息技术股份有限公司 | Behavior sequence anomaly detection method and system based on unsupervised algorithm |
CN113887674A (en) * | 2021-12-06 | 2022-01-04 | 深圳索信达数据技术有限公司 | Abnormal behavior detection method and system based on big data |
CN113887674B (en) * | 2021-12-06 | 2022-03-22 | 深圳索信达数据技术有限公司 | Abnormal behavior detection method and system based on big data |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11330009B2 (en) | Systems and methods for machine learning-based digital content clustering, digital content threat detection, and digital content threat remediation in machine learning task-oriented digital threat mitigation platform | |
CN110570244A (en) | hot-selling commodity construction method and system based on abnormal user identification | |
Inamdar et al. | Predicting cryptocurrency value using sentiment analysis | |
CN104036255A (en) | Facial expression recognition method | |
CN102291392A (en) | Hybrid intrusion detection method based on bagging algorithm | |
CN116541911B (en) | Packaging design system based on artificial intelligence | |
CN110414367A (en) | A kind of timing behavioral value method based on GAN and SSN | |
CN110134873B (en) | Microblog rumor prediction method based on forwarding network hierarchical analysis | |
CN108304479A (en) | A kind of fast density cluster double-layer network recommendation method based on graph structure filtering | |
Hua | Bitcoin price prediction using ARIMA and LSTM | |
Soni et al. | Effective machine learning approach to detect groups of fake reviewers | |
CN110913354A (en) | Short message classification method and device and electronic equipment | |
Zhang et al. | Spam detection via feature selection and decision tree | |
CN114970816A (en) | Method and device for training neural network of graph | |
CN112463848A (en) | Method, system, device and storage medium for detecting abnormal user behavior | |
Abinaya et al. | Spam detection on social media platforms | |
Shehnepoor et al. | DFraud³: multi-component fraud detection free of cold-start | |
CN112990202B (en) | Scene graph generation method and system based on sparse representation | |
Zhang et al. | Consumer credit risk assessment: A review from the state-of-the-art classification algorithms, data traits, and learning methods | |
CN116662817B (en) | Asset identification method and system of Internet of things equipment | |
Shehnepoor et al. | Spatio-temporal graph representation learning for fraudster group detection | |
US11914956B1 (en) | Unusual score generators for a neuro-linguistic behavioral recognition system | |
Vunnava et al. | Performance analysis of ml techniques in identification of fake news | |
Khanvilkar et al. | Product recommendation using sentiment analysis of reviews: a random forest approach | |
Ajmain et al. | Enhancing Sentiment Analysis using Machine Learning Predictive Models to Analyze Social Media Reviews on Junk Food |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191213 |
|
RJ01 | Rejection of invention patent application after publication |